An Improved Self-Organizing Map Method for Task Assignment and Path Planning of Multirobot in Obstacle Environment

An Improved SOM-Based Method for Multi-Robot Task Assignment and Cooperative Search in Unknown Dynamic Environments

“…The success rate of the DSOM algorithm is better than the comparison methods when the number of targets is greater than 14. In Figure 9b, when the number of targets is less than 10, the number of iterations of the comparison methods in reference [1] and [22] is less than that of the DSOM, but there is not much difference. As a comparison, the number of iterations obtained by using the DSOM is very stable and within a reasonable range.…”

“…To verify the effectiveness and superiority of the proposed method in task assignment and cooperative search under an unknown dynamic environment, comparison studies with the algorithms in reference [1] and [22] are compared in this section. The algorithm in [1] was proposed for a 3-D underwater environment.…”

“…The algorithm in [1] was proposed for a 3-D underwater environment. To avoid a contingency, thirty different environments with the same size were simulated with the comparison methods in reference [1] and [22], and the proposed method of DSOM. It should be noted that in our experiments, only the 2-D model of the algorithm in reference [1] is used; the comparison methods in reference [1] and [22] do not consider the unknown environment, and the location information of obstacles is provided to them as prior information.…”

“…The average number of iterations and the average path length of the proposed method and comparison methods in reference [1] and [19] are compared in Table 1. Compared with the comparison methods, the To further compare the performance of the comparison methods in reference [1] and [22] and DSOM, simulation experiments are performed for four cases, including a dynamic environment in which the number of targets is equal to that of the robots (Case 1), a dynamic environment in which the number of targets is more than that of the robots (Case 2), a dynamic environment in which some targets are movable (Case3), and a dynamic environment in which some obstacles are movable (Case4). For each case, the numbers and locations of the targets, robots and obstacles are the Energies 2020, 13, 3296 16 of 18 same.…”

“…A multi-robot algorithm was proposed for exploration, which is better than a depth-first search with a single robot [19]. The artificial potential field (APF) [20] is an efficient path planning algorithm [21,22]. This method has the advantage of low computational complexity, and it has long been proposed in the field of robot path planning.…”